The use of the standard hinge loss for structured outputs, for the learning to rank problem, faces two main caveats: (a) the label space, the set of all possible permutations of items to be ranked, is too large, and also less amenable to the usual dynamic-programming based techniques used for structured outputs, and (b) the supervision or training data consists of instances with multiple labels...

This paper looks into classification of documents that have hierarchical labels and are not restricted to a single label. Previous work in hierarchical classification focuses on the hierarchical perceptron (Hieron) algorithm. Hieron only supports single label learning. We investigate applying several standard multi-label learning techniques to Hieron. We then propose an extension of the algorit...

This paper introduces a geometrically inspired large-margin classifier that can be a better alternative to the Support Vector Machines (SVMs) for the classification problems with limited number of training samples. In contrast to the SVM classifier, we approximate classes with affine hulls of their class samples rather than convex hulls. For any pair of classes approximated with affine hulls, w...

We present an effective training algorithm for linearly-scored dependency parsers that implements online largemargin multi-class training (Crammer and Singer, 2003; Crammer et al., 2003) on top of efficient parsing techniques for dependency trees (Eisner, 1996). The trained parsers achieve a competitive dependency accuracy for both English and Czech with no language specific enhancements.

ABSTRACT The SNoW (Sparse Network of Winnows) ar hite ture has re ently been su essful applied to a number of natural language pro essing (NLP) problems. In this paper, we propose large margin versions of the Winnow algorithms, whi h we argue an potentially enhan e the performan e of basi Winnows (and hen e the SNoW ar hite ture). We demonstrate that the resulting methods a hieve performan e om...

In structured output learning, obtaining labeled data for real-world applications is usually costly, while unlabeled examples are available in abundance. Semi-supervised structured classification has been developed to handle large amounts of unlabeled structured data. In this work, we consider semi-supervised structural SVMs with domain constraints. The optimization problem, which in general is...

The accuracy of the k-nearest neighbor algorithm depends on the distance function used to measure similarity between instances. Methods have been proposed in the literature to learn a good distance function from a labelled training set. One such method is the large margin nearest neighbor classifier that learns a global Mahalanobis distance. We propose a mixture of such classifiers where a gati...

This work was supported by NSF grant CCF0937094 and IARPA via DoI/NBC contract number D12PC00337. The U.S. Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright annotation thereon. Disclaimer: The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official p...

Large scale systems have large numbers of inputs and outputs, and include whole chemical plants as well as some unit operations, such as paper machines, polymer film extruders, and adhesive coaters. The importance of ensuring robustness of the closed loop system to model uncertainties increases as the process dimensionality increases; hence developing algorithms for computing robustness margins...

The problem of controlling the capacity of decision trees is considered for the case where the decision nodes implement linear threshold functions. In addition to the standard early stopping and pruning procedures, we implement a strategy based on the margins of the decision boundaries at the nodes. The approach is motivated by bounds on generalization error obtained in terms of the margins of ...